Article pubs.acs.org/JAFC

Simultaneous Determination of Ethyl Carbamate and Urea in Alcoholic Beverages by High-Performance Liquid Chromatography Coupled with Fluorescence Detection Jian Zhang,*,† Guoxin Liu,† Ying Zhang,‡ Qiang Gao,† Depei Wang,† and Hao Liu*,† †

Key Laboratory of Industrial Fermentation Microbiology (Tianjin University of Science and Technology), Ministry of Education, Tianjin 300457, P. R. China ‡ Key Laboratory of Food Nutrition and Safety (Tianjin University of Science and Technology), Ministry of Education, Tianjin 300457, P. R. China S Supporting Information *

ABSTRACT: On the basis of the similar fluorescence of ethyl carbamate (EC) and urea derivatives, a high-performance liquid chromatography method coupled with fluorescence detection was developed for the simultaneous determination of EC and urea in alcoholic beverages. The chromatographic separation and derivatization conditions of EC and urea were optimized. Under the established conditions, the detection limit, relative standard deviation, linear range, and recovery were 4.8 μg/L, 1.0−4.2%, 10− 500 μg/L, and 93.8−104.6%, respectively, for EC; the corresponding values were 0.003 mg/L, 1.2−4.8%, 0.01−100 mg/L, and 90.7−104.8%, respectively, for urea. The method showed satisfactory values for precision, recovery, and sensitivity for both analytes and is well-suited for routine analysis and kinetic studies of the formation of EC from urea alcoholysis in alcoholic beverages. KEYWORDS: alcoholic beverages, ethyl carbamate, urea, HPLC−FLD

■

INTRODUCTION Ethyl carbamate (EC, urethane, NH2COOCH2CH3), a known carcinogen and mutagen, was reclassified in 2007 as probably carcinogenic to humans (Group 2A) by the International Agency for Research on Cancer (IARC).1 This compound naturally occurs in many fermented foods and alcoholic beverages at levels from nanograms to milligrams per liter.2−5 Considerable attention has been paid to EC levels in alcoholic beverages because these products make the largest contributors to EC intake by humans. Therefore, some countries, including the United States, Canada, and Brazil, have set legal limits for EC in alcoholic beverages.6−8 Urea, which is derived from raw materials or is formed as a consequence of yeast metabolism,9,10 is considered to be the main precursor for EC formation.11,12 Urea can react with ethanol to form EC with moderate kinetics at ambient temperature.13 This reaction is favored by increasing temperatures and acidic pH values.14 The content of EC is therefore higher in wines that have been stored for a long period of time and for which the temperature has not been controlled well.15 EC and urea can both be present as a result of natural formation during the brewing process. The rate of EC formation is proportional to the concentration of urea. However, EC formation can account for only an extremely small fraction of urea loss. In spite of this, the level of urea in alcoholic beverages can indicate the potential for EC formation to some extent.14,16 To obtain further kinetic information about the formation of EC from urea, both EC and urea should be quantified accurately. Many methods for detecting EC or urea in alcoholic beverages have been reported to date. For EC, various © 2014 American Chemical Society

chromatographic methods using different detectors are prevalent.17−20 For urea, three approaches have been used. These include enzymatic hydrolysis, color-forming reactions, and chromatographic separations.21−24 Although the individual detection of EC or urea has been commonly reported,17−22 to the best of our knowledge, there are very few methods29,30 capable of detecting EC and urea simultaneously. Obviously, simultaneous determination of EC and urea could reduce the system error, simplify the analysis, and improve the efficiency of the process. Xanthydrol can react with various primary amides to form xanthyl amides (Figure 1).25 Some of these amides can fluoresce and thus can be characterized by fluorescence detection (FLD) after separation by high-performance liquid chromatography (HPLC). On the basis of these findings, Herbert and co-workers developed HPLC procedures for the determination of EC17 and urea26 in alcoholic beverages. Francis et al.27 improved the previous HPLC−FLD method for urea determination with automated derivatization and optimized excitation and emission wavelengths. Mihucz et al.28 also modified the method to detect EC. These studies suggest that EC and urea could be detected simultaneously by HPLC−FLD by using the appropriate excitation and emission wavelength and suitable separation conditions. Recently, Xiong et al.29 reported the use of HPLC− FLD for the simultaneous determination of EC and urea in Received: Revised: Accepted: Published: 2797

December March 10, March 10, March 10,

3, 2013 2014 2014 2014

dx.doi.org/10.1021/jf405400y | J. Agric. Food Chem. 2014, 62, 2797−2802

Journal of Agricultural and Food Chemistry

Article

Figure 1. Formation of xanthyl amides from 9-xanthydrol and primary amides.

Figure 2. Three-dimensional fluorescence spectra of (a) XEC and (b) Xurea. Sample Derivatization. The derivatization reaction was conducted in chromatographic vials as follows: 0.4 mL of the 9-xanthydrol solution (0.02 mol of 9-xanthydrol per liter in 1-propanol) was added to 0.6 mL of a sample or standard, and then 0.1 mL of hydrochloric acid (1.5 mol/L) was slowly added. To ensure the reaction reached completion, the solution was homogenized and held for 30 min in the dark before being injected into chromatographic system for analysis. Three repeated injections of standard solutions, at six concentrations, were used to generate calibration curves. HPLC−FLD. The experiments were conducted on an Agilent 1200 series analytical HPLC system equipped with a quaternary solvent delivery system and a G1321A fluorescence detector. The separations were performed on an Agilent Zorbax Eclipse XDB-C18 column (250 mm × 4.6 mm, 5 μm). The flow rate was 1.0 mL/min, the oven temperature 30 °C, and the injection volume 20 μL. The excitation and emission wavelengths were 240 and 308 nm, respectively. Phase A was sodium acetate (pH 7.2, adjusted with acetic acid). Phase B was acetonitrile. Elution was conducted using a gradient program as indicated in Table S1 of the Supporting Information. Quantitation was performed in accordance with an external standard method. Validation. Method validation was performed following the recommendation of the International Conference on Harmonization.31 The method was validated for linearity, stability, limits of detection (LOD), limits of quantification (LOQ), precision, and accuracy.

Chinese yellow-rice wine. However, the method needed different excitation and emission wavelengths at different times. After that, Wang et al.30 developed an HPLC−FLD procedure (λex of 234 nm and λem of 600 nm) to simultaneously detect EC and urea in Chinese yellow-rice wine. The method featured high recovery and good repeatability. However, their report did not provide any spectral information about the analytes. Moreover, the understanding of the derivatization conditions is insufficient, and the detection limit for urea is too high to monitor its variation in the brewing process of some alcoholic beverages. Therefore, in this work, we further studied the HPLC−FLD method and applied it to more alcoholic beverages. First, we systematically investigated the excitation wavelength and emission wavelength of the derivatized analytes. Then, the derivatization and separation conditions of analytes were optimized. Next, the parameters of the developed method were evaluated. Finally, we applied the developed method to the analysis of real samples from the Chinese market.

■

MATERIALS AND METHODS

Reagents and Solvents. Milli-Q-RG (Millipore, Milford, MA) high-purity water was used for dissolution of chemicals and dilution. Acetic acid, ethanol, hydrochloric acid, 1-propanol, and sodium acetate, purchased from Merck (Darmstadt, Germany), were of analytical grade. Acetonitrile of HPLC grade was purchased from Merck. EC (99%) and urea (99%) were purchased from Sigma Chemical Co. (St. Louis, MO). 9-Xanthydrol (99%) was purchased from Fluka Chemicals Co. (Buchs, Switzerland). Samples. Red wines, grape brandies, and Chinese rice wines were purchased from the Chinese market. All of the information regarding the samples was obtained from the sample labels. The alcoholic strength of red wines, grape brandies, and Chinese rice wines ranged from 11.5 to 12% (v/v), 38 to 40% (v/v), and 15 to 16% (v/v), respectively. The red wines were made from Cabernet Sauvignon grapes grown in the Huaizhuo basin district (Hebei Province, P. R. China); we studied 2008, 2010, and 2011 vintages. The grape brandies were produced by three different manufacturers (namely, Greet Wall, The Dynasty, and Zhang Yu Wine Co. Ltd.) and aged for 3−15 years. The Chinese rice wines were produced by four different manufacturers (namely, Jiashan Youhua, Tapai, Gu Yue Long Shan, and Fuchunjiang Wine Co. Ltd.). All of the companies are from Zhejiang Province, P. R. China. All of the Chinese rice wines were semidry rice wines.

■

RESULTS AND DISCUSSION

Fluorescence Characteristics. To optimize fluorescence detection, 1 mg/L EC and urea in anhydrous alcohol were derivatized to obtain the three-dimensional spectra of xanthyl EC (XEC) and xanthyl urea (Xurea) (Figure 2). It is evident that XEC and the Xurea have similar fluorescence characteristics, which is consistent with the work of Francis et al.27 This similarity is caused by their similar molecular structures. The diagonal bands in panels a and b of Figure 2 could be the result of Rayleigh and Raman scattering by solvent molecules. This possibility is supported by the fact that these diagonal signals were present in the contour plot of the blank working solutions. Both XEC and Xurea exhibit maximal fluorescence at 308 nm after excitation at 213 or 240 nm. For the detection of fluorescence from Xurea, Herbert et al.26 used excitation and emission wavelengths of 233 and 600 nm, respectively. Francis et al.27 later verified that the peak at 592 2798

dx.doi.org/10.1021/jf405400y | J. Agric. Food Chem. 2014, 62, 2797−2802

Journal of Agricultural and Food Chemistry

Article

Figure 3. Determination of urea and EC in a 20% alcohol/water mixture using three gradient programs.

Figure 4. Formation of Xurea (a) and XEC (b) as a function of reaction time and temperature.

Chromatographic Separation. To achieve successful resolution of XEC and Xurea, three gradient elution programs were tested (Table S1 of the Supporting Information). Figure 3 shows a chromatogram of the derivatized urea and EC in a 20% alcohol/water mixture. An amino group (-NH2) results in Xurea being more polar than XEC, which has an ethoxy group (-OCH2CH3). This characteristic allowed easy separation of XEC and Xurea. In all cases, we observed insufficient separation of XEC. This may be due to the presence of the 9-xanthydrol derivatives in the reagent, which have polarities that are similar to that of XEC and exhibit strong fluorescence at λex and λem wavelengths of 240 and 308 nm, respectively. Alternatively, the MS detector can provide excellent selectivity for XEC and demonstrate no other interference signal.34The best analytical frequency, chromatographic resolution (not satisfactory but acceptable for XEC), and peak shape were obtained adopting gradient program II. Under this condition, XEC and Xurea are detected at 20.6 and 13.8 min, respectively. Optimization of the Derivatization Condition. The derivatization kinetics was determined by using standard solutions [1 mg/L for both EC and urea in a 40% (v/v) alcoholic solution] reacting at different temperatures. Because a higher temperature can accelerate the reaction of urea to form

nm (corresponding to 600 nm) was actually an artifact of the instrumentation. They observed similar maxima at 211 and 240 nm in the excitation spectrum and 298 nm in the emission spectrum. Our results are consistent with the report of Francis et al.;27 the slight discrepancies in the emission maxima were due (in part) to the different slit widths used in each instrument. For the detection of the fluorescence of XEC, Herbert et al.17 used excitation and emission wavelengths of 233 and 600 nm, respectively. Mihucz et al.28 later detected XEC at λex and λem wavelengths of 238 and 300 nm, respectively, because they observed that the peak at 300 nm was more intense than that at 600 nm. Our results are consistent with the report of Mihucz et al. Although wavelengths of 213 or 240 nm could excite the same intense peak at 308 nm, we chose an excitation wavelength of 240 nm and an emission wavelengths of 308 nm for the characterization of XEC and Xurea. This is because high-energy photons will be absorbed by mainly the solvent and can damage the fluorescence spectrophotometer. The use of the same excitation and emission wavelengths allowed the use of HPLC−FLD for the simultaneous determination of XEC and Xurea. 2799

dx.doi.org/10.1021/jf405400y | J. Agric. Food Chem. 2014, 62, 2797−2802

Journal of Agricultural and Food Chemistry

Article

to be more suitable for simultaneous determination of EC and urea in alcoholic beverages. This is true for the three following reasons. (1) With an increasing alcoholic strength, the response of XEC decreases sharply while that of Xurea increases mildly. Choosing a 20% (v/v) alcohol/water mixture as the solvent can significantly improve the detection sensitivity for EC but slightly decrease that of urea. (2) Alcoholic beverages generally contain EC and urea at levels of micrograms per liter and milligrams per liter, respectively. Moreover, the response of the fluorescence detector to Xurea is more significant than the response to XEC. Therefore, we must give priority to EC detection sensitivity. (3) Alcoholic beverages generally contain 10−40% alcohol. Choosing a 20% (v/v) alcohol/water mixture as the solvent results in no dilution or trace dilution of the sample, which can significantly improve the detection limits of EC and urea. Method Validation. Linearity, LOD, and LOQ. Considering the levels of EC and urea normally found in alcoholic beverages, and the maximal or recommended limits set in some countries, the linearity of EC detection was studied in the range from 10 to 500 μg/L, while the linearity of the urea detection was tested in the range from 0.01 to 100 mg/L. The regression equations used for EC and urea determination were Y = 30.5X + 6.4 and Y = 53.3X + 4.3, respectively. In the equations, Y is the peak area and X is the EC or urea concentration. The correlation coefficients (R2) of both analytes were >0.999, which indicates excellent linearity over the tested concentration ranges. The LOD (signal-to-noise ratio of 3) and LOQ (signalto-noise ratio of 10) for both analytes were as follows: 4.8 and 16 μg/L for EC and 0.003 and 0.01 mg/L for urea, respectively. Precision and Accuracy. To determine the precision and accuracy of the method, intraday and interday repeatability and recovery experiments were performed. To simplify processing, red wine, Chinese rice wine, and grape brandy were chosen for the precision and accuracy experiment because they present more complex compositions and demonstrate greater applicability of the method. Three samples were spiked with EC and urea, at low and high fortification levels, and were analyzed accordingly. The results are listed in Table 1 and show great repeatability for both analytes because the intraday RSD values were